Add support for bin, oct and hex literals

spicelang · Aug 8, 2022 · 603e302 · 603e302
1 parent d6b0fd9
commit 603e302
Show file tree

Hide file tree

Showing 5 changed files with 81 additions and 25 deletions.
diff --git a/.run/Spice_run.run.xml b/.run/Spice_run.run.xml
@@ -1,5 +1,5 @@
 <component name="ProjectRunConfigurationManager">
-  <configuration default="false" name="Spice_run" type="CMakeRunConfiguration" factoryName="Application" PROGRAM_PARAMS="run -O0 -ir ../../media/test-project/os-test.spice" REDIRECT_INPUT="false" ELEVATE="false" USE_EXTERNAL_CONSOLE="false" PASS_PARENT_ENVS_2="true" PROJECT_NAME="Spice" TARGET_NAME="Spice_run" CONFIG_NAME="Debug" RUN_TARGET_PROJECT_NAME="Spice" RUN_TARGET_NAME="Spice_run">
+  <configuration default="false" name="Spice_run" type="CMakeRunConfiguration" factoryName="Application" PROGRAM_PARAMS="run -O2 -ir ../../media/test-project/os-test.spice" REDIRECT_INPUT="false" ELEVATE="false" USE_EXTERNAL_CONSOLE="false" PASS_PARENT_ENVS_2="true" PROJECT_NAME="Spice" TARGET_NAME="Spice_run" CONFIG_NAME="Debug" RUN_TARGET_PROJECT_NAME="Spice" RUN_TARGET_NAME="Spice_run">
     <envs>
       <env name="RUN_TESTS" value="OFF" />
     </envs>

diff --git a/media/test-project/os-test.spice b/media/test-project/os-test.spice
@@ -11,11 +11,5 @@ f<int> main() {
 }*/
 
 f<int> main() {
-    int[10][10] a;
-    for int i = 0; i < 10; i++ {
-        for int j = 0; j < 10; j++ {
-            a[i][j] = i * j;
-        }
-    }
-    printf("Cell [1,3]: %d", a[1][3]);
+    printf("Test: %d\n", 0xfff);
 }
diff --git a/src/Spice.g4 b/src/Spice.g4
@@ -170,18 +170,19 @@ DOT: '.';
 ELLIPSIS: '...';
 
 // Regex tokens
+DOUBLE_LIT: [0-9]*[.][0-9]+;
+INT_LIT: NUM_LIT;
+SHORT_LIT: NUM_LIT 's';
+LONG_LIT: NUM_LIT 'l';
 CHAR_LIT: '\'' (~['\\\r\n] | '\\' (. | EOF)) '\'';
 STRING_LIT: '"' (~["\\\r\n] | '\\' (. | EOF))* '"';
-INT_LIT: NONZERO_DIGIT DIGIT* | ZERO;
-DOUBLE_LIT: DIGIT+ DOT DIGIT+;
-SHORT_LIT: INT_LIT 's';
-LONG_LIT: INT_LIT 'l';
-IDENTIFIER: NONDIGIT (NONDIGIT | DIGIT)*;
+IDENTIFIER: [a-zA-Z_][a-zA-Z0-9_]*;
 
-fragment ZERO: [0];
-fragment DIGIT: [0-9];
-fragment NONZERO_DIGIT: [1-9];
-fragment NONDIGIT: [a-zA-Z_];
+fragment NUM_LIT: DEC_LIT | BIN_LIT | HEX_LIT | OCT_LIT;
+fragment DEC_LIT: ([0][dD])?[0-9]+;
+fragment BIN_LIT: [0][bB][01]+;
+fragment HEX_LIT: [0][xXhH][0-9a-fA-F]+;
+fragment OCT_LIT: [0][oO][0-7]+;
 
 // Skipped tokens
 BLOCK_COMMENT: '/*' .*? '*/' -> skip;

diff --git a/src/parser/AstBuilderVisitor.cpp b/src/parser/AstBuilderVisitor.cpp
@@ -1197,24 +1197,21 @@ std::any AstBuilderVisitor::visitPrimitiveValue(SpiceParser::PrimitiveValueConte
       primitiveValueNode->data.doubleValue = std::stod(t->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::INT_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TY_INT;
-      primitiveValueNode->data.intValue = std::stoi(t->toString());
+      primitiveValueNode->data.intValue = parseInt(t->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree);
                t->getSymbol()->getType() == SpiceParser::SHORT_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TY_SHORT;
-      primitiveValueNode->data.shortValue = (short)std::stoi(t->toString());
+      primitiveValueNode->data.shortValue = parseShort(t->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::LONG_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TY_LONG;
-      primitiveValueNode->data.longValue = std::stoll(t->toString());
+      primitiveValueNode->data.longValue = parseLong(t->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::CHAR_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TY_CHAR;
-      primitiveValueNode->data.charValue = ctx->CHAR_LIT()->toString()[1];
+      primitiveValueNode->data.charValue = parseChar(ctx->CHAR_LIT()->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree);
                t->getSymbol()->getType() == SpiceParser::STRING_LIT) {
       primitiveValueNode->type = PrimitiveValueNode::TY_STRING;
-      std::string strValue = ctx->STRING_LIT()->toString();
-      strValue = strValue.substr(1, strValue.size() - 2);
-      replaceEscapeChars(strValue);
-      primitiveValueNode->data.stringValue = strValue;
+      primitiveValueNode->data.stringValue = parseString(ctx->STRING_LIT()->toString());
     } else if (auto t = dynamic_cast<antlr4::tree::TerminalNode *>(subTree); t->getSymbol()->getType() == SpiceParser::TRUE) {
       primitiveValueNode->type = PrimitiveValueNode::TY_BOOL;
       primitiveValueNode->data.boolValue = true;
@@ -1479,4 +1476,61 @@ void AstBuilderVisitor::replaceEscapeChars(std::string &string) {
   CommonUtil::replaceAll(string, "\\v", "\v");
   CommonUtil::replaceAll(string, "\\'", "\'");
   CommonUtil::replaceAll(string, "\\?", "\?");
+}
+
+int32_t AstBuilderVisitor::parseInt(const std::string &input) {
+  std::function<int32_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+    return std::stoi(substr, nullptr, base);
+  };
+  return parseNumeric(input, cb);
+}
+int16_t AstBuilderVisitor::parseShort(const std::string &input) {
+  std::function<int16_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+    return (int16_t)std::stoi(substr, nullptr, base);
+  };
+  return parseNumeric(input, cb);
+}
+
+int64_t AstBuilderVisitor::parseLong(const std::string &input) {
+  std::function<int64_t(const std::string &, int)> cb = [](const std::string &substr, int base) {
+    return std::stoll(substr, nullptr, base);
+  };
+  return parseNumeric(input, cb);
+}
+
+int8_t AstBuilderVisitor::parseChar(const std::string &input) { return input[1]; }
+
+std::string AstBuilderVisitor::parseString(std::string input) {
+  input = input.substr(1, input.size() - 2);
+  replaceEscapeChars(input);
+  return input;
+}
+
+template <typename T> T AstBuilderVisitor::parseNumeric(const std::string &input, std::function<T(const std::string &, int)> cb) {
+  if (input.length() >= 3) {
+    char c1 = input[0];
+    char c2 = input[1];
+    std::string substr = input.substr(2);
+    if (c1 == '0') {
+      switch (c2) {
+      case 'd':
+      case 'D':
+        return cb(substr, 10);
+      case 'b':
+      case 'B':
+        return cb(substr, 2);
+      case 'h':
+      case 'H':
+      case 'x':
+      case 'X':
+        return cb(substr, 16);
+      case 'o':
+      case 'O':
+        return cb(substr, 8);
+      default:
+        return cb(input, 10);
+      }
+    }
+  }
+  return cb(input, 10);
 }
diff --git a/src/parser/AstBuilderVisitor.h b/src/parser/AstBuilderVisitor.h
@@ -4,6 +4,7 @@
 
 #include <SpiceVisitor.h>
 
+#include <functional>
 #include <utility>
 
 // Forward declarations
@@ -86,5 +87,11 @@ class AstBuilderVisitor : public SpiceVisitor {
   std::string fileName;
 
   // Private methods
+  static int32_t parseInt(const std::string &input);
+  static int16_t parseShort(const std::string &input);
+  static int64_t parseLong(const std::string &input);
+  static int8_t parseChar(const std::string &input);
+  static std::string parseString(std::string input);
+  template <typename T> T static parseNumeric(const std::string &input, std::function<T(const std::string &, int)> cb);
   static void replaceEscapeChars(std::string &string);
 };